Wireless communication technologies, since their introduction, have evolved very quickly and people have been brought in 30 years into a much closer world. In parallel radiofrequency (RF) electromagnetic fields (EMF) are more and more used. As a consequence, people's attentions around health risks of exposure to RF EMFs have grown just as much as their usages of wireless communication technologies. Exposure to RF EMFs can be characterized using different exposure metrics (e.g., incident field metrics, absorption metrics...). However, the existing methodologies are well suited to the maximum exposure assessment for the individual under the worst-case condition. Moreover in most cases, when dealing with exposure issues, exposures linked to RF EMF emitted from base stations (BTS) and by wireless devices (e.g, mobile phones and tablets) are generally treated separately. This thesis has been dedicated to construct and validate a new method for assessing the real day-to-day RF EMF exposure to a wireless network as a whole, exploring the people's daily life, including both downlink and uplink exposures and taking into account different technologies, usages, environments, etc. Towards these objectives, we analyzed for the first time the average population exposure linked to third generation network (3G) induced EMFs, from both uplink and downlink radio emissions in different countries, geographical areas, and for different wireless device usages. Results, derived from device usage statistics, show a strong heterogeneity of exposure, both in time and space. We show that, contrary to popular belief, exposure to 3G EMFs is dominated by uplink radio emissions, resulting from voice and data traffic, and average population EMF exposure differs from one geographical area to another, as well as from one country to another, due to the different cellular network architectures and variability of mobile usage. Thus the variability and uncertainties linked to these influencing factors were characterized. And a variance-based sensitivity analysis of the global exposure was performed for the purpose of simplifying its evaluation. Finally, a substitution model was built to evaluate the day-to-day global LTE induced EMFs exposure of a population taking into account the variability linked to propagation environment, usage, as well as EMFs from personal wireless devices and BTS. Results have highlighted the importance of received power density from BTS to the issue of global exposure induced by a macro LTE network. This substitution model can be further used to analyze the evolution of the wireless network in terms of EMF exposure.